Fun with numbers

Baseball Monster is a rankings website that caters player values and projections to your league’s settings. They describe the way in which they calculate the player values as such:

“Player values are based on the standard score statistic, where a value of 0.0 is the average for your league. Anything positive is better than the league average, negative worse. As a general guideline, a value of 2.0 and above for an individual category is very good, -2.0 and below, really bad.”

Easy enough, right? The league I used for experimental purposes was a standard, mixed league with 12 teams. In the numbers below, I tweaked the player values based on numbers that I thought were realistic projections for 2012. For example, I looked at what Matt Kemp’s value might be if he had hit 25 points lower in the batting average department and, further, where he might have ranked if he only hit 30 home runs.

My findings and observations…

*

30 Had Matt Kemp hit 30 home runs rather than 39 last year, he would have still been the second best offensive player (trailing, then, only Jacoby Ellsbury). Some regression can be expected; after all, his home run rate was five and a half percent above his career average, and he never hit more than 28 home runs before his 2011 campaign.

.294 Had Matt Kemp hit 30 home runs with a .294 average (his career mark) last year, once again he would’ve found himself near the top of the value leaderboards. This time, he’d also trail Ryan Braun. Kemp won’t hit .324 again—it was aided by a whole lotta luck—but I wouldn’t say it’s the largest stretch of the imagination to assume 30+ homers and a .300 average. Don’t overestimate the regression that’ll hit him; he’s worthy of consideration at #1 overall.

*

12If Dee Gordon meets his Oliver projections of two homers, 40 RBIs, 76 runs, 40 stolen bases, and a .270 average, he would’ve been the 12th most valuable shortstop of 2011.

61 If Dee Gordon meets his Oliver projections in all categories except one—stolen bases—and beats his projection in said category by 21 steals, he’d be the seventh most valuable shortstop of 2011. I think he can steal 61 (he was on pace for that precise number with 600 plate appearances last year), and perhaps he can best his runs projection, too (on pace for 87 last year). I’d rather have him than Elvis Andrus when price is considered.

*
3.38 Heath Bell’s ERA projection from Oliver. Projected numbers of a 3.38 ERA, a 1.25 WHIP, 36 saves, three wins, and 56 strikeouts would’ve had Bell as the 50th most-valuable relief pitcher last year. Sure, the list includes a slew of middle relievers, but there are a dozen or so closers I would prefer to Bell.

3.18 Heath Bell’s career ERA away from PETCO in the last four seasons.

44 Heath Bell’s average saves total in the last three seasons. Let’s cut the man some slack.

*

.239 Evan Longoria’s BABIP last year, which led to a .244 batting average.

.319Evan Longoria’s BABIP over the previous three seasons, which led to a three year average of .283

78 Evan Longoria’s 2011 runs scored total.

A Hardball Times Update
Goodbye for now.

98 Evan Longoria’s previous two-year runs scored average.

3 Evan Longoria’s 2011 stolen base total.

12Evan Longoria’s previous two-year stolen bases average.

5 Evan Longoria’s ranks among third basemen last season, clocking in (well) behind Jose Bautista, Michael Young, Adrian Beltre, and Aramis Ramirez.

2 Evan Longoria’s retrospective ranking among third basemen with a “should have been” 31/98/99/12/.283 line.

17 Evan Longoria’s retrospective rankings among all position players with the aforementioned “should have been” line, which makes still him a “Don’t Draft,” in the first round for yours truly.

*

10 The number of position players who, per linear weights, were above average in all five major offensive categories in 2011. Their names: Matt Kemp, Jacoby Ellsbury, Ryan Braun, Dustin Pedroia, Justin Upton, Alex Gordon, Melky Cabrera, Carlos Gonzalez, Brandon Phillips, Jeff Francoeur.

0 The number of players listed above who play shortstop.

.276 League average batting average.

.273 Asdrubal Cabrera’s batting average in 2011.

4The number of categories Asdrubal Cabrera was above average in last season: home runs, stolen bases, RBIs, and runs scored.

~0.001 The percent chance that I budge and take Troy Tulowitzki in the first round, knowing that Asdrubal Cabrera can be had in the seventh (at least in our recent mock draft).

*

19 David Price’s ranks among starting pitchers in 2011, when he put up a 3.49 ERA, a 1.13 WHIP, and 218 strikeouts while he garnered only 12 wins.

14 David Price’s ranks among starting pitchers in 2011 if he had won 19 games, as he did the previous year.

11 David Price’s ranks among starting pitchers in 2011 if he had won 19 games, as he did the previous year, and if he had pitched to a 3.32 ERA, like both his FIP and xFIP suggested.

9 The number of pitchers taken before David Price in drafts, on average, per Mock Draft Central. Shame. I’d feel comfortable, still, with Price as my ace, but perhaps the early fourth round is a little early to draft a pitcher for some, especially when Ian Kennedy can be found much, much later.

*

21 The number of wins last year by Ian Kennedy. Cut the number to 16, the 2011 wins total of his Arizona counterpart, Daniel Hudson, and you still have a top eight pitcher when last year’s other stats are used. A lot of people credit Kennedy’s hefty (and likely unsustainable) win total to his value last year, but forget that he was excellent without too much luck aiding his success. His home-run rate was a tad low, so factor in a few more of those, but otherwise, draft him as an ace with confidence. But wait…

20 The number of pitchers taken ahead of Ian Kennedy in our recent mock draft.

7 The round in which Ian Kennedy was selected in our recent mock draft.

0 How much sense that makes.

*


9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
lc fiore
12 years ago

Great article. I’d love to see more like this.

Derek Ambrosino
12 years ago

I concur.

This is an awesome tool, and great for thought experimenting. Bravo, Nick. Very useful info.

MH
12 years ago

I think you put far too much weight on last year’s numbers….

15, 3.29, 1.16, 199
15, 3.24, 1.14, 200

Those are the James projections for Hudson and Kennedy, respecitively.  ZiPs has them both at a 3.42 ERA, 1.20 WHIP, and edges K/9 in favor of Kennedy by a mark of 7.88 to 7.64 (though it docks Kennedy a large number of IP for some reason).  Meanwhile Hudson is ADPing about 17 picks later, despite the fact that the projection systems think they’re about the same pitcher.  Suffice to say, I’m much more likely to have Hudson on my roster than Kennedy. 

You can do something similar with David Price and Yovani Gallardo. 

I don’t think Price is that overpriced (heh), but I do think Kennedy is at least a bit.  I don’t see him anywhere near a Top 12 SP (A #1 in a 12-team league should be a top 12 pitcher), and I’d be glad to see an opponent draft him as such.  His Whiff rate is ordinary, he’s a fly ball pitcher due for his HR/FB to rise (the difference between his 7.7% and the D’Backs team 9.5% averages is pretty significant for a fly ball guy), his LOB% will fall, and somehow his sub-90mph fastball graded as by far his best pitch last year (by FG pitch type linear weights) at an astonishing +28.7, after previously registering -7.0 over his career and coming in at -2.9 in 2010.  I see him as a fringe Top 20, more like a guy in the ~25 range and someone fairly high risk with a lot of volatility in his projection.  Yes, his BABIP matched his career rate, which is below the MLB average, but that’s still only in less than 500 IP, still isn’t that great a sample size for BABIP, so as much as a pitcher who matched his career BABIP can look like a regression candidate, IMO, Kennedy does.  I’m guessing 14, 3.60, 1.20, 175 for him.  That’s basically what I’m expecting from guys like CJ Wilson, Josh Beckett, and Shaun Marcum and less than I’m especting from Matt Garza, making him either a high end #3 or a low end #2. 

I do essentially agree with you on Kemp and Longo.

Derek Ambrosino
12 years ago

Very interesting and thorough take, MH.

I think part of the value in this column is being able to look at a player’s projections in the context of their relative fantasy value.

To say that a player’s stat line is going to regress, even offering a specific alternate stat line, is of limited value if we don’t know what a particular line of production is actually worth.

We can certainly debate the amount of regression each of us may expect from a player like Kennedy – and I don’t mean to diminish the worthiness of that debate at all – but we also need to know the relative worth of what production we settle on as a projection.

This article attempts to put projections in context, which is a step many neglect to take. Once you have your currency conversion down, you can then tweak the values based on your own opinions about the player.

Both sides of the equation are important.

MH
12 years ago

Of course, certainly agree Derek, and in that regard I do think its a terrific approach, don’t mean to disparage that at all, more the assertion that Kennedy is an “ace”. 

I think part of the issue is that a lot of people who do take that type of context into account don’t articulate it (as Nick as done) but rather leave it as an assumption.  We touch on it a little bit by throwing categorical assertions such as “Top 20” or “#3 SP” around, but those terms are vague and easily abused. 

To me, based on my impression of the trend of the run environment, a Top 12 SP should have a projection of roughly a 3.20 ERA and 1.10 WHIP or better.  FWIW, by rotochamp (since other projections on fangraphs aren’t yet sortable), the 13th best projected SP ERA is Tim Lincecum’s 3.04 (surprisingly slightly inferior to Madison Bumgarner’s 3.02). 

My relatively unscientific general rankings have the following as firmly superior to Kennedy (in no particular order):  Verlander, Halladay, Kershaw, Lee, Lincecum, Sabathia, Hamels, Weaver, Haren, Price, Greinke, Gallardo, Lester, Cain, Strasburg, Latos, Bumgarner, Garza—despite the fact that Kennedy was superior to many of these pitchers last year and his line from last year was superior to the projections of many of these pitchers.  My issue is that simply regressing a few individual aspects of a pitchers line from a season ago can give a pretty incomplete picuture, and in Kennedy’s case, there needs to be more regression than simply his Win total (although its certainly worth understanding that his line would have been elite even if you shaved off 5 wins).  Regardless, that’s 18 pitchers I would prefer to Kennedy without hesitation. 

The following group would also cause me to pause and consider the choice in a bit more detail:  Wilson, Hudson, Beckett, Hanson, Moore, Wainwright, Darvish, Anibal Sanchez, Gio Gonzalez, Luebke, and Marcum.  Kennedy would be in this group for me, though discretely where I’m not sure is particularly meaningful—the projection I place on him is similar to the projections I place on all of these guys.  Kennedy is more like Hudson, Sanchez, and Beckett in terms of volatility while Hanson, Moore, Wainwright, Darvish are riskier but carry significantly greater potential “luck-independent” reward, putting them in a different class, even if I consider their general projections similar. 

That’s 29 total pitchers.  So my guess that I would have Kennedy somewhere in the ~25ish range among pitchers in comparison to the fantasy baseball universe is quantifiably verified.  Obviously this is opinion, and again, pretty unscientific, but at least its a more specific quantification of how I value Kennedy within the context of the game.

Sam
12 years ago

MH: Very interesting take on this. Am I missing something or is Shields not in your top 30?

MH
12 years ago

My bad should have included Shields in the same group as Kennedy..

Derek Ambrosino
12 years ago

MH,

Your take is probably more in line with the consensus of the group in the mock. In the mock, Kennedy went after Cain, Wilson, Hanson (my pick), and then the next to go right after him were Garza, Moore, Latos (my pick again). Shields went next.

Will H.
12 years ago

MH: you’re very well-spoken

Nick: thanks so much for telling me about this great tool, even though I, too, am never as high on Longo as others and don’t ever think Asdrubal will far that far to me (despite this being an “expert” league). Regardless, nice analysis…